Aalborg Universitet

Theory for natural ventilation by thermal buoyancy in one zone with uniform
temperature
Terpager Andersen, K.

Published in:
Building and Environment

Publication date:
2003

Link to publication from Aalborg University

Citation for pulished version (APA):

Terpager Andersen, K. (2003). Theory for natural ventilation by thermal buoyancy in one zone with uniform
temperature. Building and Environment, 38(11), 1281-1289.

General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
You may not further distribute the material or use it for any profit-making activity or commercial gain
You may freely distribute the URL identifying the publication in the public portal ?
Take down policy
If you believe that this document breaches copyright please contact us at vbn@aub.aau.dk providing details, and we will remove access to
the work immediately and investigate your claim.

Downloaded from vbn.aau.dk on: October 08, 2014

 Available online at www.sciencedirect.com

Building and Environment 38 (2003) 1281 1289

www.elsevier.com/locate/buildenv

Theory for natural ventilation by thermal buoyancy in one zone

with uniform temperature
Karl Terpager Andersen
Energy and Indoor Climate Division, Danish Building and Urban Research, Dr. Neergaards Vej 15, Hoersholm DK-2970, Denmark
Received 19 March 2003; received in revised form 28 March 2003; accepted 6 May 2003

Abstract

Based on fundamental +ow equations describing mass balance, energy conservation and momentum, a consistent solution is derived
for natural ventilation by thermal buoyancy in a room with two openings and with uniform temperature.
The solution is a reliable tool for analysing and designing natural ventilation systems where thermal buoyancy is the dominating driving
force.
? 2003 Elsevier Ltd. All rights reserved.

Keywords: Natural ventilation; Thermal buoyancy

1. Introduction The above-mentioned models are not or only partly based

In literature, di1erent models are used for explaining the
mechanism of natural ventilation by thermal buoyancy. In
the earliest literature a column or a fan model is used. In
the column model, the driving force is assumed to be the
thermal buoyancy on a column of warm air immersed in
colder surrounding air. The column has a height equal to the
vertical distance between inlet and outlet and a cross-section
equal to the opening area of the outlet (cf. e.g. [1]). The fan
model assumes the same driving force as the column model,
and this force creates the air movement in the openings and
overcomes the friction losses.
In recent literature a third model, the pressure model,
is commonly used. It is based on the indoor and outdoor
barometric pressure distributions and it introduces the neu-
tral plane, which is the level where indoor and outdoor
pressure are equal. The position of this level is determined
by using a mass balance equation with velocities deter-
mined by the Bernoulli equation (cf. e.g. [2]). In the lat-
est literature, emptying7lling box models are introduced.
They are based on the pressure model, but include tempera-
ture strati7cation considerations known from displacement
ventilation (cf. e.g. [3]).
Tel.: +45-45-7423-88.
E-mail address: kta@dbur.dk (K.T. Andersen).
on fundamental +ow equations. Various intuitive interven-
tions or tricks are performed to establish the air+ow. Dif-
ferent calculation results occur dependent on which model
is used and how it is used, and this creates uncertainties
of analysis and design of natural ventilation systems (cf.
e.g. [3,4]).
The purpose of this paper is to set up a consistent the-
ory for natural ventilation by thermal buoyancy using the
fundamental +ow equations. Only the simplest case will be
considered in this paper. It is a room with two openings
placed at two di1erent vertical levels, with uniform indoor
temperature and with steady-state conditions.
2. Fundamental ow equations
It is assumed beforehand that the Reynolds number for the
air+ow through the openings is so high that the +ow takes
the shape of jets on entering and leaving the room. Then,
the air pressure in the smallest cross-section of the jets, the
vena contracta, is equal to the surrounding pressure and the
air velocity in these sections can be considered uniform.
The air+ow between the vena contracta of the inlet and
the vena contracta of the outlet in a room is considered, cf.
Fig. 1. This +ow does not take place in a stream tube al-
though the +ow boundaries ful7l the stream tube requirement

0360-1323/03/$ - see front matter ? 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0360-1323(03)00132-X
1282 K.T. Andersen / Building and Environment 38 (2003) 1281 1289
Nomenclature
A opening area (m2 )
Ac area of vena contracta (m2 )
Cc contraction coeHcient
Cv velocity coeHcient
H vertical distance between inlet and outlet (m)
H1 vertical distance from neutral plane to inlet (m)
H2 vertical distance from neutral plane to outlet (m)
G production rate of pollution (m3 =s)
P exchange rate of work (W)
R gas constant of air (J=(kg K))
T temperature (K)
Ti; i internal temperature by frictionless +ow (K)
c air pollution concentration (m3 =m3 )
cp speci7c heat capacity of air by constant pressure
(J=(kg K))
g gravity acceleration (m=s2 )
p pressure at the opening (Pa)
u internal energy per unit mass (J/kg)
y vertical co-ordinate of the opening (m)
vc air velocity in vena contracta (m/s)
vc; theo theoretically obtainable air velocity (i.e. by fric-
tionless +ow) (m/s)
Fig. 1. Control volume and control surface for room with two openings.
of no +uid transfer through the boundaries except through
the ends. The +uid does not follow streamlines when +ow-
ing from one vena contracta to the next. There is nor any di-
rect relationship between the +ow parameters (i.e. velocities,
cross-section areas and pressures) in lower and upper part
of the room and in the two venas contractas, like what you
wfr internal friction work per unit mass (J/kg)
Greek symbols
 net heat input (W)
0 net heat input when indoor and outdoor temper-
ature are equal (W)
E kinetic energy correction factor
 density of the gas (kg=m3 )
+ow coeHcient
 resistance coeHcient
Subscripts
acc accepted
c contracted
d discharge
i indoor
o outdoor
req required
1 inlet
2 outlet
7nd for the +ow in a stream tube. Therefore, the Bernoulli
equation is not valid for the considered air+ow. It is only
valid when the air+ow takes place in a stream tube with no
friction, no exchange of heat or work and with uniform air
velocity in any cross-section of the +ow. Even the modi7ed
version of the equation, which take friction and velocity
pro7le into account, is invalid.
2.1. Control volume
In order to ensure that all factors in+uencing the air+ow
can be taken into account, a control volume representing the
room is de7ned. The control volume of the case in ques-
tion is enclosed by the surfaces of the room, the two venas
contractas cross-sections and the two short lengths of jet be-
tween opening and vena contracta, cf. Fig. 1. In this volume
there will be no exchange of mass, energy and momentum
through the boundaries except through the two venas con-
tractas.
In a control volume, the total change of an extensive prop-
erty (i.e. a property dependent on the substance present such
as mass, energy and momentum) is considered. According to
the so-called Reynolds transport equation, the total change
is equal to the change rate of the property of the control
volume plus the eMux of the property through the control
surfaces (cf. e.g. [5,6]). In connection with the eMux, static
pressure, air velocity and cross-section areas are of interest
in the venas contractas.
 K.T. Andersen / Building and Environment 38 (2003) 1281 1289
The static pressure in the vena contracta of the inlet jet is
equal to the indoor pressure at inlet level, and in the vena
contracta of the outlet jet, the static pressure is equal to the
outdoor pressure at outlet level.
The air velocity in the venas contractas can be considered
uniform and can be determined by
vc = Cc vc; theo ;
where vc is the air velocity in vena contracta, Cv the velocity
coeHcient and vc; theo the theoretically obtainable velocity
(i.e. by frictionless +ow).
The cross-section area of a vena contracta can be
determined by
Ac = Cc A;
2
where Ac is the area of the vena contracta (m ), Cc the
contraction coeHcient and A is the opening area (m2 ).
2.2. Mass balance
For a control volume with steady-state +ow, Reynolds
transport equation results in the continuity equation, which
states that the rate of increase of mass within the control
volume plus the net eMux of mass through its control volume
is zero. For the case shown in Fig. 1, this can be expressed
by
o Ac1 vc1 i Ac2 vc2 = 0;
3
where  is the density of the gas (kg=m ), index 1 is used for
inlet, index 2 is used for outlet, index i is used for interior
conditions and index o is used for exterior conditions.
By introducing Eqs. (1) and (2) you get
o Cc1 A1 Cv1 vtheo1 i Cc2 A2 Cv2 vtheo2 = 0
or
o Cd1 A1 vtheo1 i Cd2 A2 vtheo2 = 0;
where Cd is the discharge coeHcient de7ned by
Cd = Cc Cv :
2.3. Energy conservation
Energy conservation expresses the 7rst law of thermody-
namics. For a control volume, you get an equation, which
states that the rate of change of stored energy (kinetic, po-
tential and internal energy) within the control volume plus
the eMux of stored energy across the control surface is equal
to the exchange rate of heat and work (i.e. net heat input
and exchange of power).
For the case shown in Fig. 1 with steady +ow (i.e. no
change in stored energy in the control volume), you get [5].
2
(ui2 + 12 E2 vc2 + gy2 + po2 =i )i vc2 Ac2
2
(uo1 + 12 E1 vc1 + gy1 + pi1 =o )o vc1 Ac1
= + P;
1283
where u is the internal energy per unit mass (J/kg), E the
kinetic energy correction factor, g the gravity acceleration
(m=s2 ), y the vertical co-ordinate of the opening (m), p the
pressure at the opening (Pa),  the net heat input (W) and
P the exchange rate of work (W).
The kinetic energy correction factor takes into account
that the velocity pro7le is not uniform. For a parabolic pro-
(1)
7le, you have E 1:5, and for the turbulent +ow in vena
contracta with almost uniform velocity pro7le, you have
E 1:02. In the following a value of unity is assumed.
The net heat input is the heat gain (from persons, electrical
equipment, sunshine, heating systems, etc.) minus the heat
losses (due to heat transmission through surfaces and to
(2) in7ltration).
The exchange rate of work is related to work done by,
e.g. a fan. In the case in question you have P=0. Further, you
have the exchange rate of displacement work represented by
the p=-terms.
Between internal energy, displacement work and temper-
ature you have the following relationship:
(ui; 2 + po2 =i ) (uo1 + pi1 =o ) = cp (Ti To )
=cp (Ti; i To ) + wfr ; (5a)
where Ti; i is the internal temperature by frictionless +ow
(K), cp the speci7c heat capacity of air by constant pressure
(J=(kg K)) and wfr the internal friction work per unit mass
(3a)
(J/kg).
Compared to frictionless +ow, the friction results in some-
what higher indoor temperature and a smaller air velocity.
When the friction loss takes place in an opening, it can be
expressed by
wfr = 12 vc2 ; (5b)
where  is the resistance coeHcient.
(3b) By dividing Eq. (4) with the mass +ow o vc1 Ac1 =i vc2 Ac2
and by inserting Eq. (5a) into the equation, you get with
P=0
(3c) 1 2 2
2 vc2 12 vc1 + g(y2 y1 ) + cp (Tii To )

+wfr = ; (5c)
o vc1 Ac1
where the unit for each term of the equation is J/kg.
2.3.1. Net heat input
Some heat gains and heat losses are temperature depen-
dent. For instance for increasing indoor temperature, heat
gains from persons (sensible heat) decreases and heat trans-
mission loss and in7ltration loss increases. Net heat input
can roughly be expressed by
 = 0 k(Ti To ); (6)
where 0 is net heat input when indoor and outdoor tem-
perature are equal, and k is a constant, which depends on
(4) building type and use of building.
1284 K.T. Andersen / Building and Environment 38 (2003) 1281 1289

Frequently, a 7xed indoor temperature is aimed at, so 3. Additional equations

that a temperature di1erence can be estimated with reason-
able accuracy and so that a constant net heat input can be
assumed.
2.3.2. Reduced energy equation
If considering a room with a net heat input of 4000 W,
with a vertical distance between two equally large openings
of 10 m and with a temperature di1erence of 4 K, you get an
air velocity of about 1:1 m=s in the openings and an air+ow
rate of about 0:85 m3 =s. In this case, the terms in Eq. (5c)
get the following values:
1 2
2v 0:6 J=kg;
g(y2 y1 ) 100 J=kg;
wfr 0:2 J=kg;
cp (Tii To ) 4000 J=kg;
=(o vc1 Ac1 ) 4100 J=kg:
It is seen that the net heat input and internal energy are the
totally dominating terms in Eq. (5c). The energy equation
is therefore not suitable for determining any velocities, as it
would require exact values for net heat input and temperature
di1erence. The equation is only suited for determining the
di1erence between indoor and outdoor temperatures, and it
can, with an error below 1%, be reduced to
o cp Ac1 vc1 (Ti To ) = : (7)
Of the four fundamental +ow equations, only three equa-
tions are applicable as the horizontal momentum equation is
unsuitable. The net heat input (or its constant part) is the in-
dependent variable, and there are six dependent variable, i ,
Ti , vc1 , vc2 , pi1 , and pi2 . Three more independent equations
are therefore needed. Two equations can be derived by set-
ting up fundamental +ow equations for new control volumes
in connection with inlet and outlet air+ow, respectively. The
third one can be derived from the perfect gas equation.
3.1. Outlet air:ow
For the outlet air+ow, you can consider a control volume
as shown in Fig. 2. The control surface consists of a part of
an indoor spherical surface (being entrance of the control
volume), the vena contracta section of the outlet (being the
exit), and the surface of the jet connecting vena contracta
with the spherical surface. For this control volume, you can
set up the energy equation, whereas neither the mass balance
equation nor the two momentum equations are suitable in
this case.
In the entrance of the control volume, the air velocity is
very close to zero and the static pressure is pi2 . For the exit,
the air velocity is vc2 and the pressure is po2 . Besides there
is no heat supply and height levels for entrance and exit are
equal. For this control volume, you get, when comparing
with frictionless +ow
ui2 uo1 = wfr ;

2.4. Momentum equation

Based on Newtons second law for a 7nite system, an

equation for the acting forces in a control volume can be
set up. It states that the net internal force acting on the air
in the control volume equals the time rate of change of
momentum of the air within the control volume plus the
net rate of momentum +ux or transport out of the control
volume through its surface (cf. e.g. [6]). Being a vector
equation, the momentum equation can be split into scalar
equations.
In the horizontal equation, the velocity terms are small
compared to the remaining terms as was the case in the en-
ergy equation. In combination with uncertainties on pressure
distributions, the equation is unsuitable for determining air
velocities in the openings.
In the vertical momentum equation, the vertical velocity
component in the room is negligible and the equation results
in the well-known equation for linear (hydrostatic) pressure
distribution

pi1 pi2 = i gH: (8) Fig. 2. Control volumes and control surfaces for inlet and outlet.
 K.T. Andersen / Building and Environment 38 (2003) 1281 1289 1285

which inserted into an equation similar to Eq. (4) together 3.4. Outside pressure distribution
with Eq. (5b) results in
In Eqs. (9a) and (9b), the two outside static pressures,
1 2 2
2 vc2 + po2 =i + 12 2 vc2 pi2 =i = 0 po1 at inlet level and po2 at outlet level, are introduced. As-
suming one of these pressures to be a known quantity, only
or one new dependent variable is introduced and only one new
independent equation is required. This new equation can be
2 2
pi2 po2 = Qp2 = 12 i (1 + 2 )vc2 = 12 i 2 vc2 ; (9a) the hydrostatic relationship between the two pressures, i.e.

where 2 = 1 + 2 is a constant, in the following called the
+ow coeHcient, which is solely introduced to simplify some
of the equations in following.
3.2. Inlet air:ow
For the inlet air+ow, a control volume similar to the one
for the outlet can be considered, cf. Fig. 2. In this case,
air velocity and static pressure are vc1 and po1 at entrance
and they are approximately zero and pi1 at exit. With no
heat supply and entrance and exit at same level, you get in
similarity to Eq. (9a)
2 2
po1 pi1 = Qp1 = 12 o (1 + 1 )vc1 = 12 o 1 vc1 :
po1 po2 = o gH: (11)
4. Solutions
A total of seven equations with seven dependent variables
are derived, and you can get an unambiguous solution with
net heat input as independent variable. Alternatively, the in-
door air density or the indoor air temperature can be chosen
as independent variable. For instance, the indoor air tem-
perature acts as independent variable, when this temperature
is chosen as design criteria. The result is then unambigu-
ous solutions based on air density di1erences or temperature
(9b) di1erences as shown in the following.

3.3. Density di;erence versus temperature di;erence 4.1. Solution based on density di;erences

The atmosphere can be considered as a perfect gas, where By using Eqs. (3a), (8), (9a), (9b) and (11), a solution
the relationship between pressure, density and temperature with air density di1erence as independent variable can be
is expressed by derived.

p = RT; (10a) 4.1.1. Air velocities

From Eqs. (9a) and (9b), you get
where R is the gas constant for the air. From this equation,
the following relationship between density di1erence and pi1 = po1 12 o 2
1 vc1 ; (12)
temperature di1erence can be derived:


 2
1 po pi 1 po Ti pi To pi2 = po2 + 12 o 2 vc2 : (13)
Q = o i = =
R To Ti R To Ti

 By inserting these two equations into Eq. (8), you get
po Ti (pi =po )To
= ; po1 12 o 2
(po2 + 12 o 2
RTo Ti 1 vc1 2 vc2 ) = i gH

which again, after some manipulation, can be rearranged as and further, by using Eq. (11)

 1 2
+ 12 o 2
Ti T o To po pi 2 o 1 vc1 2 vc2
Q = o 1+
Ti Ti To po = i gH + (po1 po2 ) = i gH + o gH
Ti T o
= o (1 + K): (10b) =(o i )gH = QgH: (14)
Ti
From Eq. (3a) you get
In practice, the value of K is smaller than 0:4102 . Further,
pi =po 1:0 with an error smaller than 0.02% so that i Ti vc2 = (o =i )(Ac1 =Ac2 )vc1 (15)
o T2 , cf. Eq. (10a). From Eq. (10b) you then get with an
error below 0.5% and by inserting this into Eq. (14), you get for determining

 the inlet velocity
p o T i To QT QT
Q = = o = i : (10c) 1 2 2
(Ac1 =Ac2 )2 ) = gH Q
RTo Ti Ti To 2 vc1 (o 1 + i 2 (o =i )
1286 K.T. Andersen / Building and Environment 38 (2003) 1281 1289

or

1=2
2QgH
vc1 =
o 1 + i 2 (o =i )2 (Ac1 =Ac2 )2

1=2
2QgH
=
o 1 (1 + (o =i )( 2 = 1 )(Ac1 =Ac2 )2 )

1=2
2QgH1
= ; (16)
o 1
where
H
H1 = : (17)
1 + (o =i )( 2 = 1 )(Ac1 =Ac2 )2
Similarly, you get for the outlet velocity (or by inserting Eq.
(16) into Eq. (15))

1=2
2QgH2
vc2 = ; (18)
i 2
Fig. 3. Indoor and outdoor pressure distributions for room with two
where openings and with uniform temperature.

H
H2 = : (19)
1 + (i =o )( 1 = 2 )(Ac2 =Ac1 )2 in between will be a level with neutral pressure, i.e. where
indoor and outdoor pressures are equal. This level is called
By adding Eqs. (17) and (19) you get
the neutral plane.
H H The pressure conditions are illustrated in Fig. 3. Outdoor
H 1 + H2 = + = H: (20)
1 + n 1 + 1=n pressure can be expressed by
poy = po1 o gy (23a)
4.1.2. Pressure conditions. Neutral plane
Having determined the two velocities, indoor pressures at and indoor pressure can be expressed by, when using
inlet and outlet level can be found. From Eqs. (9b) and (16) Eq. (21a)
you get
piy = pi1 i gy = po1 + QgH1 i gy: (23b)
2
pi1 = po1 12 i 1 vc1 = po1 QgH1 (21a)
4.1.3. Neutral plane position
and by using Eq. (8) you get To 7nd the neutral plane position, Eqs. (23a) and (23b)
pi2 = pi1 i gH = po1 (o i )gH1 i gH are equated, and you get
po1 o gy = po1 + (o i )gH1 i gy
or by inserting H1 = H H2 , cf. Eq. (20)
or
pi2 = po1 o gH + o gH2 + i gH i gH2 i gH
(o i )gy = (o i )gH1
= po1 o gH + o gH2 i gH2
or
= po2 + (o i )H2 : (21b)
y = H1 :
As to the pressure di1erences across the openings, you get
from Eq. (21a) for the inlet Thus, H1 is the vertical distance from neutral plane to inlet,
and from Eq. (20) it is seen that H2 is the vertical distance
Qp1 = po1 pi1 = QgH1 (22a) from neutral plane to outlet.
The two distances can be expressed in a shorter form
and for the outlet, you get from Eq. (21b)
H
H1 = ; (24)
Qp2 = pi2 po2 = QgH2 : (22b) 1 + n2

As seen, there is an indoor negative pressure at inlet level H n2 H

H2 = = ; (25)
and an indoor positive pressure at outlet level. Somewhere 1 + (1=n)2 1 + n2
 K.T. Andersen / Building and Environment 38 (2003) 1281 1289 1287

where equation for determining the velocity coeHcient Cv :

n = (o =i )( 2 = 1 )(Ac1 =Ac2 )2 2
Qp = 12 vtheo = 12 (1 + )vc2 = 12 (1 + )Cv2 vtheo
2

= (o =i )(Cd2 =Cd1 )2 (A1 =A2 )2 : (26) or

For instance for n = 1:0, you get H1 = H2 = H=2, and for 1 1
Cv = = 1=2 : (32)
n = 2:0 you get H1 = 0:2H and H2 = 0:8H . (1 + )1=2

Frequently, it can be assumed that ui and that Cd1

Cd2 . You then get
H Cc
H1 ; (27)
1 + (A1 =A2 )2
H For a simple sharp-edged opening as a window, you have
H2 : (28)
1 + (A2 =A1 )2
4.1.4. Air:ow rate
The air+ow rate through the inlet can be determined by

1=2
2QgH1
qv1 = Ac1 vc1 = Cc1 A1 Cv1 vtheo1 = Cd1 A1
o
(29)
and for the air+ow rate through the outlet, you get

1=2
2QgH2
qv2 = Ac2 vc2 = Cc2 A2 Cv2 vtheo2 = Cd2 A2 :
i
(30)
The two air+ow rates are di1erent corresponding to the two
di1erent air densities. However, the mass +ow rates through
the two openings are equal.
4.1.5. Stack e;ect and stack height
It is the quantity QgH , which creates the air+ow and
overcomes the friction in the two openings, cf. Eq. (14).
Therefore, this quantity is often called the stack e1ect, and
the vertical distance H between the openings is called the
stack height. The unit of the quantity is Pa or J=m3 . It can
be interpreted as the pressure di1erence, which creates the
air movement, or the work carried out on 1 m3 air as to give
this amount of air a certain velocity.
The pressure di1erence across the inlet and the air velocity
in the inlet depends on the vertical distance H1 between
inlet and neutral plane. For the outlet, you have a similar
dependency on the distance H2 . Therefore, the distances H1
or H2 are often called eHcient stack heights.
For the discharge coeHcient you get
Cc
Cd = Cv Cd = = 1=2 : (33)
(1 + )1=2
for instance  0:1 and Cc 0:7 so that
1
Cv 0:95
(1 + 0:1)1=2
and
Cd 0:95 0:7 0:65:
4.1.7. Air density considerations
So far, only air density di1erences due to temperature
have been considered. However, an air density di1erence
may as well be created by di1erences in moisture content.
Adding moisture to the air results in lower air density as the
weight of vapour is smaller than that of air. The relationships
between density, moisture content (or relative humidity) and
net moisture input are like the relationships between density,
temperature and net heat input.
When adding heat as well as moisture to the air, the e1ect
of the heat will be dominating under practical conditions.
Therefore, the relationship between air density di1erence
and air temperature, described by Eq. (10c), can still be
used. Only in very extreme situations will the error exceed
10%, and you will be on the safe side from a design point
of view.
4.2. Solution based on temperature di;erence
By replacing density di1erences with temperature
di1erences as expressed by Eq. (10b), you get a solution
with temperature di1erence as independent variable, and
with pressure di1erences, air velocities and air+ow rates
expressed by

4.1.6. Coe>cients o gH1 QT i gH1 QT

Qp1 = = ; (34)
By frictionless +ow, i.e.  = 0, the air velocity in an open- Ti To
ing is determined by, cf. for instance Eq. (9a) o gH2 QT i gH2 QT
Qp2 = = ; (35)
vtheo = (2Qp=) 1=2
(31) Ti To

1=2
and you get 2gH1 QT
vc1 = ; (36)
2 1 Ti
Qp = 12 vtheo :

1=2
In a more general case with friction, you get from an equa- 2gH2 QT
vc2 = ; (37)
tion similar to Eq. (9a), and by using Eq. (1), the following 2 To
1288 K.T. Andersen / Building and Environment 38 (2003) 1281 1289

1=2 so that you get

2gH1 QT
qv1 = Cd1 A1 ; (38)
2=3
1=3
Ti  1
QT = 7:1 105 Ti : (44)

1=2 Cd1 A1 H1
2gH2 QT
qv2 = Cd2 A2 ; (39) When using the outlet conditions, you similarly get
To

2=3
1=3
where 5  1
QT = 7:5 10 To : (45)
Cd2 A2 H2
H
H1 = ; (40) By replacing temperature di1erences with net heat input
1 + (Ti =To )(Cd1 =Cd2 )(A1 =A2 )2
according to Eqs. (44) and (45), you get a solution with net
H heat input as independent variable, and with air velocities
H2 = : (41)
1 + (To =Ti )(Cd2 =Cd1 )(A2 =A1 )2 and air+ow rates expressed by

1=3
1=2
H1 1
4.3. Solution based on net heat input vc1 = 0:037 ; (46)
Cd1 A1 1

1=3
1=2
The di1erence between indoor and outdoor temperature is H2 1
due to the net heat input, and from Eq. (7) you get following vc2 = 0:038 ; (47)
Cd2 A2 2
relationship between the two quantities:
1=3
1=3 2=3 1
   qv1 = Cc1 A1 vc1 = 0:037(H1 ) A1
QT = = = : (42) Cd1
o cp Ac1 vc1 o cp qV1 i cp qV2
Cc1
However, the air+ow rate in the denominator is not a 7xed 1=2
= 0:037(H1 )1=3 (Cd1 A1 )2=3 (48)
quantity. As seen from Eq. (38) it is dependent on the tem- 1
perature di1erence, i.e. as Cc1 = 11=2 = Cd1 , cf. Eq. (32).
 Similarly, you get the following air+ow rate through the
QT = : (43)
o cp Cd1 A1 ((2gH1 QT )=Ti )1=2 outlet:

For a 7xed net heat input, you get a temperature di1erence qV2 = 0:038(H2 )1=3 (Cd2 A2 )2=3 : (49)
that results in an air+ow rate, which removes an amount of
heat per time unit equal to the net heat input. 4.3.2. Temperature-dependent net heat input
If net heat input is temperature dependent as shown by
4.3.1. Constant net heat input Eq. (6), you get the following equation instead of Eq. (44):
Assuming constant net heat input, Eq. (43) yields, when QT = K(0 kQT )2=3
solving it with regard to QT

2=3
1=3 or
 1
QT = Ti : QT 3 = K 3 (0 kQT )2 : (50)
o Ti cp Cd1 A1 2gH1
Outdoor air density can be replaced by outdoor temperature Solving this equation of third degree, you get a complicated
by using Eq. (10a), and after some manipulation you get expression for QT . It can be solved iteratively.

2=3
1=3
2=3
2=3
1=3
R 1 To  1 4.4. Design calculations
QT = Ti :
p o cp 2g Ti Cd1 A1 H1
In the design situation, the task is to calculate the re-
For the constants, following values can be inserted: quired opening areas to ensure acceptable indoor air quality
R = 287 J=(kg K); in winter and acceptable indoor thermal comfort in summer.
In winter as well as in summer, a required air+ow rate is
po = 101 300 Pa; determined. In winter, when acceptable indoor air quality
is represented by an acceptable CO2 or moisture level, the
cp = 1010 J=(kg K); required air+ow rate is determined by
g = 9:82 m=s2 : G
qVreq = ; (51)
(ci; acc co )
Further, under practical conditions you have with an error
below 2% where G is the production rate of pollution (m3 =s), ci; acc the
accepted indoor pollution concentration (m3 =m3 ) and co is
(To =Ti )2=3 = (1 QT=Ti )2=3 0:97; the outdoor pollution concentration (m3 =m3 ).
 K.T. Andersen / Building and Environment 38 (2003) 1281 1289 1289

In summer, when acceptable thermal comfort is repre- be taken with this fundamental method as to get correct
sented by an acceptable air temperature Ti; acc , the required solutions in case of indoor temperature strati7cation or if
air+ow rate is determined by, cf. Eq. (7) the room has more than two openings above each other.
 Likewise it is possible to analyse analytically such questions
qV; req = : (52) of importance for design and control as
cp o (Ti; acc T0 )
In the design situation, known quantities are building geom- Is the opening orientation (vertical, horizontal or sloped)
etry, represented by the vertical opening distance H, and the of any importance?
opening area rate n = A1 =A2 . What happens, when the neutral plane intersects an open-
In cases where it is preferable to work with a solution ing, and how can this be avoided?
based on temperature di1erence, the required inlet area is What is the optimal opening area ratio?
determined by, cf. Eq. (38) What is the relationship between air+ow rate, opening
qV; req
A1; req = : (53) area and opening area ratio when either temperature dif-
Cd1 ((2gH1 QT )=Ti )1=2 ference or net heat input is kept constant?
If the solution is based on net heat input, you get from
Eq. (48) By using the fundamental method, a reliable tool is obtained
3=2 for analysing and designing natural ventilation systems in
qV; req
A1 = 140 : (54) situations where thermal buoyancy is the dominating driving
Cd1 (H1 )1=2 force.
By solving Eq. (44), accepted indoor temperature can be
included, and you get

1=2
3=2 References
7  1 Ti
A1; req = 6:0 10 : (55)
Cd1 H1 QTacc [1] Bruce JM. Natural ventilation by stack e1ect. Farm Building Progress
From the required inlet area, the required outlet area can be 1973;32:238.
[2] Bruce JM. Natural convection through openings and its application
found as A2; req = A1; req =n. to cattle building ventilation. Journal of Agricultural Engineering
Research 1978;23:15167.
[3] Li Y. Buoyancy-driven natural ventilation in a thermally one-zone
5. Conclusion building. Building and Environment 2000;35:20714.
[4] Foster MP, Down MJ. Ventilation of livestock buildings by natural
By setting up and solving the fundamental +ow equations, convection. Journal of Agricultural Engineering Research 1987;37:
113.
you get a solution with all parameters taken into account. [5] Shames IH. Mechanics of +uids. New York: McGraw-Hill, 1992.
This gives a consistent explanation of the mechanism of [6] Olson RM, Wright SJ. Essentials of engineering +uid mechanics. New
natural ventilation by thermal buoyancy. Further steps can York: Harper & Row, 1990.